BePro

Achieving Better Benefits via Flexible Feature Matching in Post-Deduplication Delta Compression （IPDPS'2025）

Some of the code references Deepsketch and Fastcdc, the Paper and source code are as follows:

• DeepSketch: a new machine learning-based reference search technique for post-deduplication delta compression （FAST’2022）
  • Source Code: https://github.com/dgist-datalab/deepsketch-fast2022
• FastCDC: A Fast and Efficient Content-Defined Chunking Approach for Data Deduplication (Usenix ATC'2016)
  • Source Code: https://github.com/wxiacode/FastCDC-c

Dataset

Scilab: https://github.com/scilab/scilab/tags

LNX: https://github.com/torvalds/linux/tags

TAR: http://ftp.gnu.org/gnu/

VM: https://www.osboxes.org/vmware-images/

SOF: https://www.brentozar.com/archive/2015/10/how-to-download-the-stack-overflow-database-via-bittorrent/

SYN: 10 synthetic backups of SOF_2010 by simulating file addition/deletion/modify